UCASpatial

Ultra-precision Deconvolution of Spatial Transcriptomics

Unveiling Spatiotemporal Cellular Dynamics in Complex Microenvironments

Get Started View Tutorial Sample Datasets

Key Features

Ultra-precision

Improves precision of mapping cell subpopulations to spatial locations

Entropy-based Weighting

Leverages gene contribution through entropy-based weighting for cell identity

Robust & Accurate

Enhanced identification of low-abundant and transcriptionally heterogeneous cell subpopulations

Multi-platform

Support for both R and Python, including 10X Visium HD datasets

Quick Start

R

# Install
devtools::install_github('https://github.com/BIGHanLab/UCASpatial/')

# Load data
sc_ref <- readRDS("path/to/sc_reference.rds")  # cluster info in 'sc.ref$celltype'
st_vis <- readRDS("path/to/st_data.rds")

# Run UCASpatial
result <- UCASpatial_deconv(sc_ref, st_vis, clust_vr="celltype")

Python

# Import
import UCASpatial_ds_R1

# Load data
sc_ref = sc.read("path/to/sc_reference.h5ad")
st_vis = sc.read("path/to/st_data.h5ad")

# Run UCASpatial
ucas = UCASpatial_ds_R1.UCASpatial(sc_ref, st_vis, clust_vr='celltype')
result = ucas.run()